52 



THE AMERICAN NATURALIST 



[Vol. LV 



number that would mean optimum conditions for none and doubt 

 concerning the reliability of all. 



The main problem in connection with the environment (2) is 

 to find a kind of food that will allow full development of even 

 very weak classes. It is in this field that the greatest changes in 

 method have been made. For some years — from 1910 to 1916 — 

 some modification of the fermented^banana method of preparing 

 food was followed. Ripe sound bananas were peeled, and the 

 pulp left for about 24 hours in a liquid containing yeast. This 

 liquid was usually the fermented juice from the previous lot of 

 bananas. About 25 grams of this fermented banana was put in 

 the bottom of a culture bottle and covered with absorbent paper. 



It was suspected that the real food of the larvae was not pri- 

 marily the banana but was rather the yeast cells and perhaps 

 also the bacteria, the banana being mainly the culture medium 

 for the yeast. This has been established by the work of Northrop,^ 

 of Loeb and Northrop,^ and of Baumberger.^ In July, 1916, in 

 consultation with Northrop, I started experiments with a view to 

 using as a culture medium standardized solutions, instead of 

 banana. The solution was absorbed and held in a cake in the 

 bottom of the culture bottle by shredded paper toweling, which 

 offered extensive surface for the growth of the yeast. This 

 method was unsuccessful ; the flies laid few eggs and these were 

 often overgrown by the yeast and killed. Esters and other chem- 

 icals with fruit odors did not lead to greater egg production. 

 Perhaps better success with culture solutions would be obtained 

 in supplementing and modifying banana methods. 



The banana method was modified with a view to discouraging 

 the growth of moulds and putrefactive bacteria by mild anti- 

 septics or correctives, such as benzoate, thymol, formaldehyde, 

 alcohol, powdered marble for neutralization of excessive acidity, 

 etc. Good results were obtained with alcohol, where several ex- 

 tensive sets of comparative tests seemed to show that about 1.5 

 per cent, of alcohol in the food was desirable. The most success- 

 ful alcohol method was roughly as follows: The pulp of sound 

 ripe bananas was weighed and put with an equal number of c.c. 

 of 3 per cent, alcohol in a shallow, covered dish. No yeast was 

 added, since enough wild yeast was usually present. The food 



1 Jour. Biol. Chem., 1917, pp. 181-187. 



